Skip to content

Commit 564d272

Browse files
vecxozvecxoz
committed
Maintenance for tests (LogisticRegression default arguments)
1 parent d41c4af commit 564d272

File tree

4 files changed

+148
-148
lines changed

4 files changed

+148
-148
lines changed

‎tests/test_func_api_classification_binary.py‎

Lines changed: 39 additions & 39 deletions
Original file line numberDiff line numberDiff line change
@@ -83,13 +83,13 @@ def tearDown(self):
8383

8484
def test_oof_pred_mode(self):
8585

86-
model = LogisticRegression(random_state=0)
86+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
8787
S_train_1 = cross_val_predict(model, X_train, y = y_train, cv = n_folds,
8888
n_jobs = 1, verbose = 0, method = 'predict').reshape(-1, 1)
8989
_ = model.fit(X_train, y_train)
9090
S_test_1 = model.predict(X_test).reshape(-1, 1)
9191

92-
models = [LogisticRegression(random_state=0)]
92+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')]
9393
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
9494
regression = False, n_folds = n_folds, shuffle = False, save_dir=temp_dir,
9595
mode = 'oof_pred', random_state = 0, verbose = 0, stratified = True)
@@ -110,12 +110,12 @@ def test_oof_pred_mode(self):
110110

111111
def test_oof_mode(self):
112112

113-
model = LogisticRegression(random_state=0)
113+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
114114
S_train_1 = cross_val_predict(model, X_train, y = y_train, cv = n_folds,
115115
n_jobs = 1, verbose = 0, method = 'predict').reshape(-1, 1)
116116
S_test_1 = None
117117

118-
models = [LogisticRegression(random_state=0)]
118+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')]
119119
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
120120
regression = False, n_folds = n_folds, shuffle = False, save_dir=temp_dir,
121121
mode = 'oof', random_state = 0, verbose = 0, stratified = True)
@@ -136,12 +136,12 @@ def test_oof_mode(self):
136136

137137
def test_pred_mode(self):
138138

139-
model = LogisticRegression(random_state=0)
139+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
140140
S_train_1 = None
141141
_ = model.fit(X_train, y_train)
142142
S_test_1 = model.predict(X_test).reshape(-1, 1)
143143

144-
models = [LogisticRegression(random_state=0)]
144+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')]
145145
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
146146
regression = False, n_folds = n_folds, shuffle = False, save_dir=temp_dir,
147147
mode = 'pred', random_state = 0, verbose = 0, stratified = True)
@@ -171,16 +171,16 @@ def test_oof_pred_bag_mode(self):
171171
y_tr = y_train[tr_index]
172172
X_te = X_train[te_index]
173173
y_te = y_train[te_index]
174-
model = LogisticRegression(random_state=0)
174+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
175175
_ = model.fit(X_tr, y_tr)
176176
S_test_temp[:, fold_counter] = model.predict(X_test)
177177
S_test_1 = st.mode(S_test_temp, axis = 1)[0]
178178

179-
model = LogisticRegression(random_state=0)
179+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
180180
S_train_1 = cross_val_predict(model, X_train, y = y_train, cv = n_folds,
181181
n_jobs = 1, verbose = 0, method = 'predict').reshape(-1, 1)
182182

183-
models = [LogisticRegression(random_state=0)]
183+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')]
184184
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
185185
regression = False, n_folds = n_folds, shuffle = False, save_dir=temp_dir,
186186
mode = 'oof_pred_bag', random_state = 0, verbose = 0, stratified = True)
@@ -210,14 +210,14 @@ def test_pred_bag_mode(self):
210210
y_tr = y_train[tr_index]
211211
X_te = X_train[te_index]
212212
y_te = y_train[te_index]
213-
model = LogisticRegression(random_state=0)
213+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
214214
_ = model.fit(X_tr, y_tr)
215215
S_test_temp[:, fold_counter] = model.predict(X_test)
216216
S_test_1 = st.mode(S_test_temp, axis = 1)[0]
217217

218218
S_train_1 = None
219219

220-
models = [LogisticRegression(random_state=0)]
220+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')]
221221
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
222222
regression = False, n_folds = n_folds, shuffle = False, save_dir=temp_dir,
223223
mode = 'pred_bag', random_state = 0, verbose = 0, stratified = True)
@@ -242,13 +242,13 @@ def test_pred_bag_mode(self):
242242

243243
def test_oof_pred_mode_proba(self):
244244

245-
model = LogisticRegression(random_state=0)
245+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
246246
S_train_1 = cross_val_predict(model, X_train, y = y_train, cv = n_folds,
247247
n_jobs = 1, verbose = 0, method = 'predict_proba')
248248
_ = model.fit(X_train, y_train)
249249
S_test_1 = model.predict_proba(X_test)
250250

251-
models = [LogisticRegression(random_state=0)]
251+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')]
252252
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
253253
regression = False, n_folds = n_folds, shuffle = False, stratified = True,
254254
mode = 'oof_pred', random_state = 0, verbose = 0, needs_proba = True, save_dir=temp_dir)
@@ -269,12 +269,12 @@ def test_oof_pred_mode_proba(self):
269269

270270
def test_oof_mode_proba(self):
271271

272-
model = LogisticRegression(random_state=0)
272+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
273273
S_train_1 = cross_val_predict(model, X_train, y = y_train, cv = n_folds,
274274
n_jobs = 1, verbose = 0, method = 'predict_proba')
275275
S_test_1 = None
276276

277-
models = [LogisticRegression(random_state=0)]
277+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')]
278278
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
279279
regression = False, n_folds = n_folds, shuffle = False, stratified = True,
280280
mode = 'oof', random_state = 0, verbose = 0, needs_proba = True, save_dir=temp_dir)
@@ -295,12 +295,12 @@ def test_oof_mode_proba(self):
295295

296296
def test_pred_mode_proba(self):
297297

298-
model = LogisticRegression(random_state=0)
298+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
299299
S_train_1 = None
300300
_ = model.fit(X_train, y_train)
301301
S_test_1 = model.predict_proba(X_test)
302302

303-
models = [LogisticRegression(random_state=0)]
303+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')]
304304
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
305305
regression = False, n_folds = n_folds, shuffle = False, stratified = True,
306306
mode = 'pred', random_state = 0, verbose = 0, needs_proba = True, save_dir=temp_dir)
@@ -331,18 +331,18 @@ def test_oof_pred_bag_mode_proba(self):
331331
y_tr = y_train[tr_index]
332332
X_te = X_train[te_index]
333333
y_te = y_train[te_index]
334-
model = LogisticRegression(random_state=0)
334+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
335335
_ = model.fit(X_tr, y_tr)
336336
col_slice_fold = slice(fold_counter * n_classes, fold_counter * n_classes + n_classes)
337337
S_test_temp[:, col_slice_fold] = model.predict_proba(X_test)
338338
for class_id in range(n_classes):
339339
S_test_1[:, class_id] = np.mean(S_test_temp[:, class_id::n_classes], axis = 1)
340340

341-
model = LogisticRegression(random_state=0)
341+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
342342
S_train_1 = cross_val_predict(model, X_train, y = y_train, cv = n_folds,
343343
n_jobs = 1, verbose = 0, method = 'predict_proba')
344344

345-
models = [LogisticRegression(random_state=0)]
345+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')]
346346
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
347347
regression = False, n_folds = n_folds, shuffle = False, save_dir=temp_dir,
348348
mode = 'oof_pred_bag', random_state = 0, verbose = 0, stratified = True, needs_proba = True)
@@ -382,7 +382,7 @@ def test_pred_bag_mode_proba(self):
382382
y_tr = y_train[tr_index]
383383
X_te = X_train[te_index]
384384
y_te = y_train[te_index]
385-
model = LogisticRegression(random_state=0)
385+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
386386
_ = model.fit(X_tr, y_tr)
387387
col_slice_fold = slice(fold_counter * n_classes, fold_counter * n_classes + n_classes)
388388
S_test_temp[:, col_slice_fold] = model.predict_proba(X_test)
@@ -391,7 +391,7 @@ def test_pred_bag_mode_proba(self):
391391

392392
S_train_1 = None
393393

394-
models = [LogisticRegression(random_state=0)]
394+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')]
395395
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
396396
regression = False, n_folds = n_folds, shuffle = False, save_dir=temp_dir,
397397
mode = 'pred_bag', random_state = 0, verbose = 0, stratified = True, needs_proba = True)
@@ -425,17 +425,17 @@ def test_oof_pred_bag_mode_shuffle(self):
425425
y_tr = y_train[tr_index]
426426
X_te = X_train[te_index]
427427
y_te = y_train[te_index]
428-
model = LogisticRegression(random_state=0)
428+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
429429
_ = model.fit(X_tr, y_tr)
430430
S_test_temp[:, fold_counter] = model.predict(X_test)
431431
S_test_1 = st.mode(S_test_temp, axis = 1)[0]
432432

433-
model = LogisticRegression(random_state=0)
433+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
434434
# !!! Important. Here we pass CV-generator not number of folds <cv = kf>
435435
S_train_1 = cross_val_predict(model, X_train, y = y_train, cv = kf,
436436
n_jobs = 1, verbose = 0, method = 'predict').reshape(-1, 1)
437437

438-
models = [LogisticRegression(random_state=0)]
438+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')]
439439
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
440440
regression = False, n_folds = n_folds, shuffle = True, save_dir=temp_dir,
441441
mode = 'oof_pred_bag', random_state = 0, verbose = 0, stratified = True)
@@ -462,15 +462,15 @@ def test_oof_pred_bag_mode_shuffle(self):
462462
#---------------------------------------------------------------------------
463463
def test_oof_mode_metric(self):
464464

465-
model = LogisticRegression(random_state=0)
465+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
466466
scorer = make_scorer(accuracy_score)
467467
scores = cross_val_score(model, X_train, y = y_train, cv = n_folds,
468468
scoring = scorer, n_jobs = 1, verbose = 0)
469469
mean_str_1 = '%.8f' % np.mean(scores)
470470
std_str_1 = '%.8f' % np.std(scores)
471471

472472

473-
models = [LogisticRegression(random_state=0)]
473+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')]
474474
S_train, S_test = stacking(models, X_train, y_train, X_test,
475475
regression = False, n_folds = n_folds, save_dir=temp_dir,
476476
mode = 'oof', random_state = 0, verbose = 0, stratified = True)
@@ -499,15 +499,15 @@ def test_oof_mode_metric(self):
499499
#---------------------------------------------------------------------------
500500
def test_oof_mode_metric_proba(self):
501501

502-
model = LogisticRegression(random_state=0)
502+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
503503
scorer = make_scorer(log_loss, needs_proba = True)
504504
scores = cross_val_score(model, X_train, y = y_train, cv = n_folds,
505505
scoring = scorer, n_jobs = 1, verbose = 0)
506506
mean_str_1 = '%.8f' % np.mean(scores)
507507
std_str_1 = '%.8f' % np.std(scores)
508508

509509

510-
models = [LogisticRegression(random_state=0)]
510+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')]
511511
S_train, S_test = stacking(models, X_train, y_train, X_test,
512512
regression = False, n_folds = n_folds, save_dir=temp_dir,
513513
mode = 'oof', random_state = 0, verbose = 0, stratified = True,
@@ -536,7 +536,7 @@ def test_oof_mode_metric_proba(self):
536536
def test_oof_pred_mode_2_models(self):
537537

538538
# Model a
539-
model = LogisticRegression(random_state=0)
539+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
540540
S_train_1_a = cross_val_predict(model, X_train, y = y_train, cv = n_folds,
541541
n_jobs = 1, verbose = 0, method = 'predict').reshape(-1, 1)
542542
_ = model.fit(X_train, y_train)
@@ -552,7 +552,7 @@ def test_oof_pred_mode_2_models(self):
552552
S_train_1 = np.c_[S_train_1_a, S_train_1_b]
553553
S_test_1 = np.c_[S_test_1_a, S_test_1_b]
554554

555-
models = [LogisticRegression(random_state=0),
555+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr'),
556556
GaussianNB()]
557557
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
558558
regression = False, n_folds = n_folds, shuffle = False, save_dir=temp_dir,
@@ -584,12 +584,12 @@ def test_oof_pred_bag_mode_2_models(self):
584584
y_tr = y_train[tr_index]
585585
X_te = X_train[te_index]
586586
y_te = y_train[te_index]
587-
model = LogisticRegression(random_state=0)
587+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
588588
_ = model.fit(X_tr, y_tr)
589589
S_test_temp[:, fold_counter] = model.predict(X_test)
590590
S_test_1_a = st.mode(S_test_temp, axis = 1)[0]
591591

592-
model = LogisticRegression(random_state=0)
592+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
593593
S_train_1_a = cross_val_predict(model, X_train, y = y_train, cv = n_folds,
594594
n_jobs = 1, verbose = 0, method = 'predict').reshape(-1, 1)
595595

@@ -615,7 +615,7 @@ def test_oof_pred_bag_mode_2_models(self):
615615
S_train_1 = np.c_[S_train_1_a, S_train_1_b]
616616
S_test_1 = np.c_[S_test_1_a, S_test_1_b]
617617

618-
models = [LogisticRegression(random_state=0),
618+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr'),
619619
GaussianNB()]
620620
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
621621
regression = False, n_folds = n_folds, shuffle = False, save_dir=temp_dir,
@@ -639,7 +639,7 @@ def test_oof_pred_bag_mode_2_models(self):
639639
def test_oof_pred_mode_proba_2_models(self):
640640

641641
# Model a
642-
model = LogisticRegression(random_state=0)
642+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
643643
S_train_1_a = cross_val_predict(model, X_train, y = y_train, cv = n_folds,
644644
n_jobs = 1, verbose = 0, method = 'predict_proba')
645645
_ = model.fit(X_train, y_train)
@@ -655,7 +655,7 @@ def test_oof_pred_mode_proba_2_models(self):
655655
S_train_1 = np.c_[S_train_1_a, S_train_1_b]
656656
S_test_1 = np.c_[S_test_1_a, S_test_1_b]
657657

658-
models = [LogisticRegression(random_state=0),
658+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr'),
659659
GaussianNB()]
660660
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
661661
regression = False, n_folds = n_folds, shuffle = False, stratified = True,
@@ -689,14 +689,14 @@ def test_oof_pred_bag_mode_proba_2_models(self):
689689
y_tr = y_train[tr_index]
690690
X_te = X_train[te_index]
691691
y_te = y_train[te_index]
692-
model = LogisticRegression(random_state=0)
692+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
693693
_ = model.fit(X_tr, y_tr)
694694
col_slice_fold = slice(fold_counter * n_classes, fold_counter * n_classes + n_classes)
695695
S_test_temp[:, col_slice_fold] = model.predict_proba(X_test)
696696
for class_id in range(n_classes):
697697
S_test_1_a[:, class_id] = np.mean(S_test_temp[:, class_id::n_classes], axis = 1)
698698

699-
model = LogisticRegression(random_state=0)
699+
model = LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr')
700700
S_train_1_a = cross_val_predict(model, X_train, y = y_train, cv = n_folds,
701701
n_jobs = 1, verbose = 0, method = 'predict_proba')
702702

@@ -727,7 +727,7 @@ def test_oof_pred_bag_mode_proba_2_models(self):
727727

728728

729729

730-
models = [LogisticRegression(random_state=0),
730+
models = [LogisticRegression(random_state=0, solver='liblinear', multi_class='ovr'),
731731
GaussianNB()]
732732
S_train_2, S_test_2 = stacking(models, X_train, y_train, X_test,
733733
regression = False, n_folds = n_folds, shuffle = False, save_dir=temp_dir,

0 commit comments

Comments
 (0)